INDEX

Explanations

instances where things are divided into distinct parts or groups

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "split" and related terms that denote division or separation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

agara

-0.75

trak

-0.68

za

-0.68

ties

-0.65

cycles

-0.65

ECH

-0.64

jamin

-0.62

oken

-0.62

instein

-0.60

oha

-0.60

POSITIVE LOGITS

 hairs

1.22

creen

0.99

 evenly

0.93

 between

0.91

 opinion

0.74

 Between

0.73

 Personality

0.70

SPL

0.68

owship

0.68

ters

0.68

Activations Density 0.083%

instances where things are divided into distinct parts or groups

occurrences of the word "split" and related terms that denote division or separation

No Comments

No Known Activations