INDEX

Explanations

crime-related events such as shootings, killings, injuries, and police responses

oai_token-act-pair · gpt-3.5-turbo

references to violent incidents or casualties

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

 fulfil

-0.69

 ç¥ŀ

-0.69

 philosophers

-0.67

papers

-0.66

pse

-0.66

 intellectuals

-0.64

 prag

-0.64

 guaranteeing

-0.64

 univers

-0.63

 Tolkien

-0.63

POSITIVE LOGITS

 RCMP

0.85

 Investigators

0.84

 NYPD

0.83

Police

0.83

police

0.80

 detectives

0.80

 LAPD

0.80

 juveniles

0.78

 police

0.78

 Police

0.77

Activations Density 1.683%

crime-related events such as shootings, killings, injuries, and police responses

references to violent incidents or casualties

No Comments

No Known Activations