INDEX

Explanations

phrases related to technological activities or tools

oai_token-act-pair · gpt-3.5-turbo

verbs and nouns related to action and manipulation

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

ãĥĹ

-0.65

º

-0.60

capacity

-0.60

WR

-0.58

arsity

-0.57

immer

-0.57

aida

-0.57

ãĤ´ãĥ³

-0.56

 constellation

-0.56

ãĥ¼ãĥĨãĤ£

-0.56

POSITIVE LOGITS

 oneself

0.93

ulate

0.90

ify

0.89

 them

0.89

 ourselves

0.86

 everything

0.86

 yourself

0.85

edly

0.85

 THEM

0.83

him

0.82

Activations Density 0.352%

phrases related to technological activities or tools

verbs and nouns related to action and manipulation

No Comments

No Known Activations