INDEX

Explanations

occupations or titles, often followed by specific context or descriptions

oai_token-act-pair · gpt-3.5-turbo

instances of "for" followed by various subjects or contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

tom

-0.67

————————————————

-0.63

cheat

-0.62

Thumbnail

-0.61

ril

-0.60

amaru

-0.60

venants

-0.59

rome

-0.58

pex

-0.57

rote

-0.57

POSITIVE LOGITS

 sake

1.19

 purposes

1.12

 reasons

0.97

erning

0.72

icion

0.66

 lovers

0.65

ummies

0.63

 reason

0.63

 adventurous

0.61

 enthusiast

0.60

Activations Density 0.292%

occupations or titles, often followed by specific context or descriptions

instances of "for" followed by various subjects or contexts

No Comments

No Known Activations