INDEX

Explanations

adjectives describing a change, improvement, or comparison over time

oai_token-act-pair · gpt-3.5-turbo

references to past experiences or comparisons to previous states

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

 intervening

-0.72

FTC

-0.65

DragonMagazine

-0.64

geries

-0.63

 reinforcements

-0.62

 premature

-0.58

ortunately

-0.58

 Engineers

-0.58

 Warm

-0.58

 Serbian

-0.58

POSITIVE LOGITS

learn

0.83

 deem

0.82

 despise

0.81

 dreamed

0.79

 deems

0.78

cra

0.77

 inhabit

0.76

 aspire

0.75

 dearly

0.73

 termed

0.73

Activations Density 0.130%

adjectives describing a change, improvement, or comparison over time

references to past experiences or comparisons to previous states

No Comments

No Known Activations