INDEX

Explanations

numbers related to measurements or quantities in various contexts

oai_token-act-pair · gpt-3.5-turbo

numerical values associated with specific categories or measurements

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pocket

-0.75

bard

-0.75

chal

-0.73

hol

-0.72

erald

-0.71

ben

-0.71

 Chronicle

-0.70

hunt

-0.69

roma

-0.69

host

-0.68

POSITIVE LOGITS

 ILCS

0.79

inyl

0.78

 Seym

0.74

 msec

0.69

 neighb

0.67

 Palestin

0.67

 sqor

0.66

 swaps

0.66

 Feet

0.65

iatus

0.65

Activations Density 0.032%

numbers related to measurements or quantities in various contexts

numerical values associated with specific categories or measurements

No Comments

No Known Activations