INDEX

Explanations

explicit answers or responses

oai_token-act-pair · gpt-3.5-turbo

instances of questions being answered

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

zinski

-0.73

gotten

-0.72

ammy

-0.70

wana

-0.70

bats

-0.70

ovie

-0.70

robat

-0.69

erker

-0.69

chin

-0.67

heric

-0.67

POSITIVE LOGITS

ysis

1.19

 answer

1.15

answer

1.14

 answ

1.05

swers

0.99

 Answer

0.97

Answer

0.96

 answering

0.94

 answered

0.91

 Answers

0.84

Activations Density 0.016%

explicit answers or responses

instances of questions being answered

No Comments

No Known Activations