INDEX

Explanations

news events and their reporting

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "reported" as it relates to news articles

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

venge

-0.84

 gauge

-0.75

 palate

-0.72

knife

-0.72

isable

-0.71

otin

-0.70

animate

-0.66

ogy

-0.66

ophone

-0.66

Gau

-0.66

POSITIVE LOGITS

reports

0.76

 quoting

0.74

 Politico

0.74

 outlets

0.73

 sources

0.72

 reports

0.71

 Newsweek

0.70

 headlines

0.69

 ARTICLE

0.68

 headlined

0.68

Activations Density 0.054%

news events and their reporting

occurrences of the term "reported" as it relates to news articles

No Comments

No Known Activations