INDEX

Explanations

mentions of high-ranking legal and government positions

oai_token-act-pair · gpt-3.5-turbo

titles and designations related to legal and political positions

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

NetMessage

-0.81

idable

-0.64

iland

-0.63

tten

-0.61

 gamble

-0.61

rero

-0.60

urat

-0.60

kinson

-0.59

iasm

-0.58

 mortals

-0.58

POSITIVE LOGITS

Ake

0.72

Sed

0.67

ector

0.65

vironments

0.65

geon

0.64

incial

0.64

 Enforcement

0.64

endment

0.64

ooter

0.63

uties

0.63

Activations Density 0.096%

mentions of high-ranking legal and government positions

titles and designations related to legal and political positions

No Comments

No Known Activations