INDEX

Explanations

interactions between individuals involving attempted coercion or force

oai_token-act-pair · gpt-3.5-turbo

actions associated with confrontational or aggressive encounters involving law enforcement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 inval

-0.94

 weddings

-0.78

 Tolkien

-0.77

 nowadays

-0.76

 pilgr

-0.74

 orphans

-0.72

 honoured

-0.71

obyl

-0.70

 castles

-0.69

 bloggers

-0.69

POSITIVE LOGITS

 handcuffs

1.01

 handcuffed

0.89

 handc

0.85

 restrain

0.83

CPR

0.82

 restraints

0.80

 handgun

0.78

raints

0.76

 restrained

0.74

roc

0.74

Activations Density 0.436%

interactions between individuals involving attempted coercion or force

actions associated with confrontational or aggressive encounters involving law enforcement

No Comments

No Known Activations