INDEX

Explanations

medical terms and condition-related phrases

oai_token-act-pair · gpt-3.5-turbo

clinical terms related to psychological and medical conditions

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

unia

-0.61

 ware

-0.58

geries

-0.57

 compan

-0.56

 bucket

-0.55

anooga

-0.55

 booted

-0.55

ãĥĹ

-0.54

rite

-0.53

 unemploy

-0.53

POSITIVE LOGITS

],[

1.14

–

1.13

â

1.11

âĢ¢âĢ¢

0.96

âĢĲ

0.90

 âĩ

0.90

reviewed

0.88

Î²

0.85

âĪĴ

0.81

].

0.80

Activations Density 0.058%

medical terms and condition-related phrases

clinical terms related to psychological and medical conditions

No Comments

No Known Activations