INDEX

Explanations

terms related to actions or decisions that have significant impact or are newsworthy

oai_token-act-pair · gpt-3.5-turbo

instances of the word "move" as it relates to actions or decisions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

omial

-0.73

 Barton

-0.70

oola

-0.70

icum

-0.67

sung

-0.65

culus

-0.65

rome

-0.64

 Divinity

-0.64

vana

-0.64

oured

-0.62

POSITIVE LOGITS

Motion

0.90

able

0.84

brates

0.80

Move

0.75

move

0.75

backs

0.74

itures

0.71

ments

0.71

 toward

0.71

 forward

0.71

Activations Density 0.036%

terms related to actions or decisions that have significant impact or are newsworthy

instances of the word "move" as it relates to actions or decisions

No Comments

No Known Activations