INDEX

Explanations

descriptions of criminal incidents involving law enforcement officers

oai_token-act-pair · gpt-3.5-turbo

phrases indicating incidents or events involving police and violence

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

Æ

-0.82

ographs

-0.82

Rules

-0.81

alties

-0.81

casts

-0.80

iments

-0.79

ãĤ¨

-0.78

faces

-0.77

views

-0.77

ATURES

-0.76

POSITIVE LOGITS

 burglary

1.10

 bystand

1.06

 disturbance

1.05

 woman

1.04

 policeman

1.02

 gunman

0.99

 classmate

0.97

 confrontation

0.97

 colleague

0.96

 stray

0.95

Activations Density 0.476%

descriptions of criminal incidents involving law enforcement officers

phrases indicating incidents or events involving police and violence

No Comments

No Known Activations