INDEX

Explanations

phrases related to actions and consequences

oai_token-act-pair · gpt-3.5-turbo

references to people and their actions or roles within various contexts

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

 CrossRef

-0.63

rition

-0.57

iven

-0.55

atever

-0.54

ASY

-0.54

again

-0.54

 Vish

-0.52

ulkan

-0.52

UNCH

-0.52

ERAL

-0.52

POSITIVE LOGITS

pires

0.67

 fitt

0.66

 paycheck

0.64

 swear

0.61

 Kardash

0.59

 fitness

0.59

 fart

0.58

 backgrounds

0.56

 themselves

0.55

liv

0.55

Activations Density 1.244%

phrases related to actions and consequences

references to people and their actions or roles within various contexts

No Comments

No Known Activations