INDEX

Explanations

mentions of actions being done involuntarily or under pressure

oai_token-act-pair · gpt-3.5-turbo

instances of being compelled or coerced into actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

vironment

-0.81

ership

-0.80

issance

-0.73

mbuds

-0.68

amara

-0.66

awks

-0.65

aghd

-0.64

ergy

-0.63

razen

-0.62

 Excellence

-0.62

POSITIVE LOGITS

 overtime

0.82

 laborers

0.80

otom

0.79

untarily

0.73

 cooker

0.73

 aback

0.72

 shut

0.69

 labou

0.69

ierrez

0.68

 confront

0.68

Activations Density 0.031%

mentions of actions being done involuntarily or under pressure

instances of being compelled or coerced into actions

No Comments

No Known Activations