INDEX

Explanations

words related to medical conditions and health issues, specifically focusing on symptoms, treatments, and outcomes

oai_token-act-pair · gpt-3.5-turbo

phrases related to health crises and medical emergencies

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 unison

-0.81

 yourselves

-0.80

 apiece

-0.73

 Their

-0.69

arity

-0.65

 seeded

-0.65

their

-0.65

ahime

-0.64

 Massacre

-0.63

 merger

-0.62

POSITIVE LOGITS

 medication

1.13

 medications

1.03

 himself

0.97

 herself

0.89

his

0.89

 therapy

0.84

 chemotherapy

0.84

 hallucinations

0.82

 belongings

0.82

 medicines

0.81

Activations Density 0.529%

words related to medical conditions and health issues, specifically focusing on symptoms, treatments, and outcomes

phrases related to health crises and medical emergencies

No Comments

No Known Activations