INDEX

Explanations

approval and disapproval ratings related to political figures

oai_token-act-pair · gpt-3.5-turbo

terms related to approval ratings and public opinion

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

Mek

-0.78

utical

-0.75

 Rout

-0.69

udic

-0.69

GV

-0.69

unct

-0.68

orpor

-0.66

aer

-0.66

SourceFile

-0.66

 Herrera

-0.66

POSITIVE LOGITS

 ratings

1.01

 rating

0.93

 polled

0.86

 Gallup

0.82

 undecided

0.81

 majorities

0.80

rated

0.79

HRC

0.79

rating

0.77

 Ratings

0.76

Activations Density 0.098%

approval and disapproval ratings related to political figures

terms related to approval ratings and public opinion

No Comments

No Known Activations