INDEX

Explanations

words related to strong emotions or characteristics

oai_token-act-pair · gpt-3.5-turbo

the occurrences of the term "res" with various contexts and relevance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

CPR

-0.70

Tet

-0.67

 Eagle

-0.67

onomy

-0.64

 Cruiser

-0.63

uyomi

-0.63

 Icelandic

-0.62

zza

-0.62

 Holmes

-0.62

 Ballard

-0.61

POSITIVE LOGITS

pec

1.03

ourced

1.02

res

1.00

uable

0.96

 cryst

0.96

earchers

0.95

olute

0.90

 unden

0.87

ailability

0.87

olver

0.87

Activations Density 0.004%

words related to strong emotions or characteristics

the occurrences of the term "res" with various contexts and relevance

No Comments

No Known Activations