INDEX

Explanations

occurrences of the word "news" or related terms

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

mentions of news agencies or sources

oai_token-act-pair · gpt-3.5-turbo

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

İĭ

-0.75

istically

-0.74

istic

-0.73

staking

-0.71

iasis

-0.67

 rapists

-0.66

uras

-0.64

 communism

-0.64

hesis

-0.63

 rapist

-0.60

POSITIVE LOGITS

åº

0.99

chool

0.96

atcher

0.94

ystem

0.93

ashington

0.90

hower

0.87

ynski

0.87

velt

0.83

ource

0.82

ea

0.82

Activations Density 0.023%

occurrences of the word "news" or related terms

mentions of news agencies or sources

No Comments

No Known Activations