INDEX

Explanations

mention of the term "Top" followed by a number, potentially relating to rankings or lists

oai_token-act-pair · gpt-3.5-turbo

references to rankings or top lists

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

ATIONS

-0.58

 Lauder

-0.57

 daylight

-0.55

 Schwe

-0.54

ities

-0.54

ÂłÂłÂł

-0.53

 Judd

-0.53

Mub

-0.52

ITED

-0.52

 foreseeable

-0.52

POSITIVE LOGITS

ographical

1.31

ography

1.21

ographic

1.17

ographically

1.15

iaries

1.13

iary

1.10

eka

1.08

ographies

1.06

 notch

1.05

ology

1.00

Activations Density 0.043%

mention of the term "Top" followed by a number, potentially relating to rankings or lists

references to rankings or top lists

No Comments

No Known Activations