INDEX

Explanations

conversation patterns with questions and answers

oai_token-act-pair · gpt-3.5-turbo

specific responses or affirmations within a conversation

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

 blot

-0.71

ded

-0.64

 undone

-0.63

Os

-0.62

 targets

-0.62

 uninterrupted

-0.60

 visible

-0.59

foreseen

-0.59

 Dems

-0.58

 airborne

-0.56

POSITIVE LOGITS

Answer

1.19

 Answer

1.07

 answer

0.97

maxwell

0.96

 Nope

0.88

swer

0.85

ingham

0.82

aber

0.82

swers

0.81

 answ

0.79

Activations Density 0.357%

conversation patterns with questions and answers

specific responses or affirmations within a conversation

No Comments

No Known Activations