INDEX

Explanations

references or attributions

oai_token-act-pair · gpt-3.5-turbo

references to specific sources or citations in a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Poles

-0.72

 Ashe

-0.63

SPONSORED

-0.62

worms

-0.59

ctica

-0.59

spo

-0.57

 Cout

-0.57

CHO

-0.57

################

-0.57

 exceptions

-0.57

POSITIVE LOGITS

actor

1.28

raction

1.26

ractive

1.22

ract

1.21

ocused

1.20

ashion

1.18

inished

1.13

eree

1.13

utation

1.13

ocusing

1.11

Activations Density 0.029%

references or attributions

references to specific sources or citations in a text

No Comments

No Known Activations