INDEX

Explanations

phrases related to titles or labels

oai_token-act-pair · gpt-3.5-turbo

references to popular cultural events or discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

crop

-0.68

TOR

-0.64

udeau

-0.61

 Shows

-0.61

ecake

-0.61

onto

-0.61

atra

-0.59

commit

-0.59

puter

-0.59

abiding

-0.58

POSITIVE LOGITS

 term

2.08

 euphem

1.90

 terminology

1.77

 phrase

1.68

 word

1.59

 adjective

1.57

 moniker

1.56

 acronym

1.54

 terms

1.50

 adject

1.48

Activations Density 1.005%

phrases related to titles or labels

references to popular cultural events or discussions

No Comments

No Known Activations