INDEX

Explanations

phrases related to police response and emergencies

oai_token-act-pair · gpt-3.5-turbo

references to law enforcement and crime-related incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Tokens

-0.85

 veto

-0.74

puff

-0.71

TPP

-0.68

 endif

-0.68

oos

-0.68

 spons

-0.65

gob

-0.64

è¦

-0.64

 folds

-0.64

POSITIVE LOGITS

 disturbance

1.02

 disturbances

0.92

 Incident

0.87

 gunfire

0.79

 Spokane

0.76

 paramedics

0.76

 suspicions

0.76

 suspic

0.75

 emergency

0.74

 complaints

0.74

Activations Density 0.374%

phrases related to police response and emergencies

references to law enforcement and crime-related incidents

No Comments

No Known Activations