INDEX

Explanations

profession-related terms and associated contexts like professional standards, professional athletes, ratings, professional journalism, and ransomware groups

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

edin

-0.88

anus

-0.74

apo

-0.72

worms

-0.71

attr

-0.69

ource

-0.68

pty

-0.68

_>

-0.67

erella

-0.67

aukee

-0.66

POSITIVE LOGITS

 wrestler

1.11

 athlete

1.02

 footballer

1.02

 wrestling

0.98

 athletes

0.97

gol

0.93

digy

0.92

ising

0.89

ized

0.88

 sports

0.86

Activations Density 0.034%

profession-related terms and associated contexts like professional standards, professional athletes, ratings, professional journalism, and ransomware groups

No Comments

No Known Activations