INDEX

Explanations

phrases related to apologies

oai_token-act-pair · gpt-3.5-turbo

instances of apologies and related phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

edin

-0.88

ulhu

-0.85

arkin

-0.80

retty

-0.78

inson

-0.78

kefeller

-0.78

enson

-0.77

eele

-0.75

olson

-0.74

KEY

-0.74

POSITIVE LOGITS

 behalf

0.94

 missing

0.80

 unres

0.78

 sins

0.70

 cance

0.69

 wrongdoing

0.69

 departed

0.69

 coward

0.67

 Cambodia

0.65

 Mara

0.65

Activations Density 0.165%

phrases related to apologies

instances of apologies and related phrases

No Comments

No Known Activations