INDEX

Explanations

controversial topics or subjects

oai_token-act-pair · gpt-3.5-turbo

references to controversial subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

abiding

-0.85

vation

-0.83

abetic

-0.77

thia

-0.75

nings

-0.75

ILA

-0.74

á

-0.73

abet

-0.72

ruary

-0.71

united

-0.71

POSITIVE LOGITS

 topic

0.99

 topics

0.95

 aspects

0.93

 proposition

0.91

 aspect

0.90

 propositions

0.89

 decisions

0.88

 opinions

0.86

 remarks

0.84

 controversial

0.84

Activations Density 0.073%

controversial topics or subjects

references to controversial subjects

No Comments

No Known Activations