INDEX

Explanations

linguistic structures related to the cause and effect relationship, particularly focusing on the outcome or result

oai_token-act-pair · gpt-3.5-turbo

phrases indicating outcomes or results

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

cill

-0.62

 unequ

-0.62

NetMessage

-0.58

ustomed

-0.56

thur

-0.55

selling

-0.54

rete

-0.53

UCT

-0.53

zag

-0.53

Finish

-0.53

POSITIVE LOGITS

is

1.11

 consists

1.08

 involves

1.07

 includes

1.04

 varies

1.01

 seems

1.00

 revolves

0.97

 appears

0.96

isn

0.93

 depends

0.93

Activations Density 0.234%

linguistic structures related to the cause and effect relationship, particularly focusing on the outcome or result

phrases indicating outcomes or results

No Comments

No Known Activations