INDEX

Explanations

questions starting with "What are" or "What were"

oai_token-act-pair · gpt-3.5-turbo

questions starting with "What are" and concepts related to inquiry or seeking information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

odder

-0.77

owan

-0.76

cul

-0.73

entity

-0.69

culated

-0.68

amphetamine

-0.67

oire

-0.67

aneously

-0.64

Hunt

-0.64

matter

-0.63

POSITIVE LOGITS

 these

0.90

 they

0.86

 some

0.85

we

0.84

you

0.84

 THEY

0.82

YOU

0.81

 those

0.79

 THESE

0.79

 SOME

0.75

Activations Density 0.034%

questions starting with "What are" or "What were"

questions starting with "What are" and concepts related to inquiry or seeking information

No Comments

No Known Activations