INDEX

Explanations

mentions of suspects or individuals related to criminal activities

oai_token-act-pair · gpt-3.5-turbo

references to suspects in criminal contexts

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

ammy

-0.80

psey

-0.77

unch

-0.77

abeth

-0.77

split

-0.73

cribed

-0.72

ummer

-0.72

Morning

-0.72

nz

-0.70

tem

-0.69

POSITIVE LOGITS

 suspect

1.07

 suspects

1.06

 perpetrator

0.81

 mishand

0.81

 perpetrators

0.73

 suspected

0.72

ively

0.69

 warrants

0.69

CSI

0.69

 offenders

0.69

Activations Density 0.018%

mentions of suspects or individuals related to criminal activities

references to suspects in criminal contexts

No Comments

No Known Activations