INDEX

Explanations

phrases related to requesting further engagement or action from the reader

oai_token-act-pair · gpt-3.5-turbo

references to additional content or newsletters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

liest

-0.62

glers

-0.61

tein

-0.57

rou

-0.56

amide

-0.56

atan

-0.55

 unto

-0.55

 rift

-0.54

tal

-0.54

witz

-0.54

POSITIVE LOGITS

Entry

0.63

stories

0.61

doing

0.61

NESS

0.59

 newsletters

0.58

adas

0.55

ire

0.54

vertis

0.54

 importantly

0.53

ilda

0.53

Activations Density 0.025%

phrases related to requesting further engagement or action from the reader

references to additional content or newsletters

No Comments

No Known Activations