INDEX

Explanations

references to large scale events or objects

oai_token-act-pair · gpt-3.5-turbo

instances of the word "massive."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

yer

-0.78

wich

-0.77

comes

-0.75

ople

-0.75

tein

-0.75

clair

-0.72

alias

-0.72

keeper

-0.72

Dialogue

-0.70

eem

-0.70

POSITIVE LOGITS

 earthqu

1.12

 amounts

0.98

 overhaul

0.95

 influx

0.93

ulously

0.88

 disproportion

0.86

 proportions

0.84

 swath

0.83

 conglomer

0.81

 spike

0.80

Activations Density 0.015%

references to large scale events or objects

instances of the word "massive."

No Comments

No Known Activations