INDEX

Explanations

phrases related to being secured or fastened to something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tymology

-0.72

TPPStreamerBot

-0.71

 reprinted

-0.70

anooga

-0.68

nder

-0.64

arted

-0.62

employed

-0.59

awei

-0.59

quickShipAvailable

-0.59

DEN

-0.59

POSITIVE LOGITS

 ensure

0.97

ggles

0.94

 compensate

0.93

 accommodate

0.90

 facilitate

0.86

 minimize

0.86

 coincide

0.85

 avoid

0.84

 accompany

0.84

 conserve

0.84

Activations Density 0.215%

phrases related to being secured or fastened to something

instances of the word "to" in various contexts

No Comments

No Known Activations