INDEX

Explanations

political and controversial terms

oai_token-act-pair · gpt-3.5-turbo

mentions of political affiliations and extremism

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ften

-0.65

uden

-0.63

mma

-0.62

 sailed

-0.60

orah

-0.59

hid

-0.57

ologue

-0.57

 smelled

-0.57

acus

-0.57

 Introduced

-0.56

POSITIVE LOGITS

 situations

0.83

 replacements

0.83

 positions

0.80

 versions

0.80

 scenarios

0.79

 backgrounds

0.78

 activities

0.78

 endorsements

0.77

 connections

0.76

 pursuits

0.76

Activations Density 0.978%

political and controversial terms

mentions of political affiliations and extremism

No Comments

No Known Activations