INDEX

Explanations

references to advertisement or promotional content

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "ads"

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

hold

-0.64

 communion

-0.61

 needle

-0.61

 Myers

-0.60

SIZE

-0.59

 Kardashian

-0.58

 snowy

-0.57

 condol

-0.57

ebin

-0.56

vette

-0.56

POSITIVE LOGITS

igmatic

1.05

ync

0.97

venture

0.95

oras

0.95

itionally

0.93

uce

0.89

ules

0.89

kins

0.88

olph

0.88

enza

0.86

Activations Density 0.011%

references to advertisement or promotional content

occurrences of the word "ads"

No Comments

No Known Activations