INDEX

Explanations

mentions of illegal activities

oai_token-act-pair · gpt-3.5-turbo

terms related to illegal activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

enger

-0.81

pread

-0.81

oleon

-0.80

ivities

-0.79

rike

-0.75

etics

-0.75

phasis

-0.74

igating

-0.72

issance

-0.72

=-=-=-=-

-0.72

POSITIVE LOGITS

 illegal

1.02

 trafficking

0.92

 downloading

0.86

 immigrant

0.86

 illegally

0.82

 immigrants

0.81

 aliens

0.81

 alien

0.81

 substances

0.78

 unfocusedRange

0.77

Activations Density 0.026%

mentions of illegal activities

terms related to illegal activities

No Comments

No Known Activations