INDEX

Explanations

terms related to political entities, specifically the term 'bloc'

oai_token-act-pair · gpt-3.5-turbo

references to political or economic alliances and groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ctory

-0.72

sbm

-0.67

agate

-0.66

Wheel

-0.65

cyclop

-0.63

Educ

-0.63

ipples

-0.62

Adv

-0.62

OSE

-0.61

vironments

-0.61

POSITIVE LOGITS

 bloc

1.25

 affili

0.82

ratulations

0.77

 tremend

0.76

atu

0.74

 membership

0.72

 grouping

0.71

pes

0.71

 elector

0.70

lin

0.70

Activations Density 0.005%

terms related to political entities, specifically the term 'bloc'

references to political or economic alliances and groups

No Comments

No Known Activations