INDEX

Explanations

names and related terms including "Max"

oai_token-act-pair · gpt-3.5-turbo

references to the name "Max."

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

Äĩ

-0.63

GROUND

-0.58

selves

-0.57

 tack

-0.57

ATIONS

-0.57

AMES

-0.56

 revolving

-0.56

ARY

-0.56

ATIONAL

-0.56

FORE

-0.56

POSITIVE LOGITS

imus

1.58

imil

1.48

imize

1.44

imal

1.26

ima

1.18

imo

1.10

ims

1.08

ime

1.07

imum

0.97

imen

0.96

Activations Density 0.038%

names and related terms including "Max"

references to the name "Max."

No Comments

No Known Activations