INDEX

Explanations

names of individuals involved in various activities or events

oai_token-act-pair · gpt-3.5-turbo

names associated with individuals and their relationships

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

etheless

-0.65

 redistributed

-0.58

 cytok

-0.56

 referen

-0.56

displayText

-0.54

 added

-0.54

 migrated

-0.53

 sped

-0.52

 logical

-0.52

 deduction

-0.51

POSITIVE LOGITS

dan

0.62

zi

0.60

 Productions

0.59

aka

0.57

 Syndrome

0.57

 Photography

0.56

arde

0.56

Jr

0.56

aux

0.55

idas

0.55

Activations Density 0.812%

names of individuals involved in various activities or events

names associated with individuals and their relationships

No Comments

No Known Activations