INDEX

Explanations

references to the term "Quinn."

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Quinn."

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

xual

-0.77

¶æ

-0.72

 fuels

-0.70

 welf

-0.69

 marrow

-0.69

ãĤ·ãĥ£

-0.64

 Arabs

-0.64

pts

-0.63

ĻĤ

-0.63

ampunk

-0.61

POSITIVE LOGITS

spir

0.95

lich

0.92

 Quinn

0.91

ipeg

0.88

pei

0.80

gren

0.79

stown

0.77

ite

0.75

ifest

0.73

sburgh

0.72

Activations Density 0.015%

references to the term "Quinn."

mentions of the name "Quinn."

No Comments

No Known Activations