INDEX

Explanations

commands or directives to perform an action

oai_token-act-pair · gpt-3.5-turbo

calls to action related to finding or discovering something

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

perty

-0.68

 fracturing

-0.67

qqa

-0.65

ossession

-0.63

forming

-0.60

endiary

-0.60

adding

-0.60

ħĭ

-0.57

WARNING

-0.57

slow

-0.57

POSITIVE LOGITS

 yourself

0.96

ById

0.93

out

0.93

 fault

0.86

 clues

0.84

 answers

0.83

 loopholes

0.79

articles

0.78

abilia

0.78

 info

0.75

Activations Density 0.060%

commands or directives to perform an action

calls to action related to finding or discovering something

No Comments

No Known Activations