INDEX

Explanations

instructional phrases indicating tasks or actions that need to be performed

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing necessity or requirements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ahime

-0.96

ustomed

-0.61

 antit

-0.61

uum

-0.61

 unequ

-0.59

 diam

-0.57

 disproportion

-0.57

Yas

-0.56

[|

-0.56

 disappro

-0.56

POSITIVE LOGITS

 save

0.77

save

0.70

rous

0.68

 besides

0.68

 except

0.67

ibaba

0.64

soever

0.61

imum

0.61

CU

0.59

except

0.59

Activations Density 0.091%

instructional phrases indicating tasks or actions that need to be performed

phrases emphasizing necessity or requirements

No Comments

No Known Activations