INDEX

Explanations

phrases related to statements or quotations

oai_token-act-pair · gpt-3.5-turbo

colons followed by subsequent content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 tremend

-0.85

ilion

-0.77

 misunder

-0.75

ĸļ

-0.72

 activ

-0.72

 trouble

-0.68

wre

-0.67

 respawn

-0.67

 queue

-0.66

llular

-0.66

POSITIVE LOGITS

 Provided

1.08

 Cosponsors

0.82

Who

0.77

 Exactly

0.76

 Between

0.76

 "âĢ¦

0.75

 They

0.75

"$

0.75

"'

0.74

 Scores

0.74

Activations Density 0.151%

phrases related to statements or quotations

colons followed by subsequent content

No Comments

No Known Activations