INDEX

Explanations

proper nouns or specific names of entities

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate something is widely recognized or identified

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

cel

-0.84

otom

-0.80

prus

-0.78

otos

-0.76

oton

-0.76

plet

-0.74

ŃĶ

-0.73

adies

-0.73

lot

-0.73

ingers

-0.72

POSITIVE LOGITS

 known

0.97

 Known

0.96

NESS

0.87

KNOWN

0.84

ledged

0.78

comings

0.76

Known

0.75

=]

0.75

Ð»

0.74

known

0.72

Activations Density 0.025%

proper nouns or specific names of entities

phrases that indicate something is widely recognized or identified

No Comments

No Known Activations