INDEX

Explanations

proper nouns and terms related to current events and media personalities

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and their roles within a political context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 mathemat

-0.80

 mould

-0.78

 glim

-0.76

 recognise

-0.72

 exha

-0.71

 envis

-0.70

 realise

-0.69

[&

-0.68

obyl

-0.67

 licences

-0.67

POSITIVE LOGITS

 tweeted

1.14

 retweet

1.10

TMZ

1.09

 Politico

1.03

 MSNBC

0.98

NBC

0.96

 Hannity

0.96

CNN

0.96

RNC

0.95

 tweet

0.91

Activations Density 1.731%

proper nouns and terms related to current events and media personalities

references to specific individuals and their roles within a political context

No Comments

No Known Activations