INDEX

Explanations

mentions or references to political affiliations, particularly the term "liberals"

oai_token-act-pair · gpt-3.5-turbo

references to liberals and their interactions or relationships within political contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ogh

-0.69

 Rune

-0.69

aum

-0.66

ardo

-0.66

iary

-0.65

 Bram

-0.65

rone

-0.64

 scouting

-0.64

ushima

-0.63

tree

-0.62

POSITIVE LOGITS

 liberals

3.29

 Liberals

2.62

 progressives

2.20

 leftists

2.18

 liberalism

2.02

 libertarians

1.92

 feminists

1.80

 conservatives

1.71

 socialists

1.67

 Democrats

1.67

Activations Density 0.032%

mentions or references to political affiliations, particularly the term "liberals"

references to liberals and their interactions or relationships within political contexts

No Comments

No Known Activations