INDEX

Explanations

organizations or entities that are being reviewed or evaluated

oai_token-act-pair · gpt-3.5-turbo

instances of the word "review" in various contexts

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

kt

-0.76

Corn

-0.73

 Hots

-0.72

âĹ¼

-0.71

pox

-0.62

Known

-0.61

risome

-0.61

FL

-0.61

rier

-0.60

Æ

-0.59

POSITIVE LOGITS

opian

0.96

 process

0.89

spective

0.86

 review

0.78

 judgment

0.76

 feasibility

0.75

 criteria

0.74

oga

0.73

rators

0.73

etts

0.73

Activations Density 0.024%

organizations or entities that are being reviewed or evaluated

instances of the word "review" in various contexts

No Comments

No Known Activations