INDEX

Explanations

terms related to law enforcement or legal professions

oai_token-act-pair · gpt-3.5-turbo

references to law and legal terminology

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

bite

-0.77

hang

-0.71

Hidden

-0.71

LV

-0.71

Newsletter

-0.69

Ò

-0.69

URA

-0.67

RS

-0.65

iron

-0.65

Course

-0.64

POSITIVE LOGITS

chio

0.90

enza

0.89

ity

0.84

raft

0.82

 acid

0.81

iance

0.76

 acids

0.76

ulty

0.75

ities

0.73

iant

0.73

Activations Density 0.025%

terms related to law enforcement or legal professions

references to law and legal terminology

No Comments

No Known Activations