INDEX

Explanations

references to reporters or journalists

oai_token-act-pair · gpt-3.5-turbo

references to journalists or individuals identified as reporters in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ensions

-0.93

ateral

-0.79

emonic

-0.75

perm

-0.74

elsius

-0.74

inence

-0.73

ension

-0.73

icted

-0.70

¥ŀ

-0.70

arching

-0.70

POSITIVE LOGITS

 reporter

0.96

 journalist

0.82

 Journalists

0.79

 Reporter

0.78

ially

0.78

 laureate

0.76

 writer

0.76

 interviewer

0.74

 columnist

0.74

 COVER

0.74

Activations Density 0.034%

references to reporters or journalists

references to journalists or individuals identified as reporters in the text

No Comments

No Known Activations