INDEX

Explanations

terms related to a person's history or past achievements, specifically focusing on their track record

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 pand

-0.72

 Krug

-0.65

LY

-0.63

 Lauder

-0.61

 Cheryl

-0.61

vernment

-0.59

ately

-0.58

 Correspond

-0.57

 Brav

-0.57

 Alvarez

-0.56

POSITIVE LOGITS

suit

1.32

suits

1.18

record

1.16

pad

1.14

tracks

1.08

track

1.06

balls

0.95

ball

0.93

masters

0.93

marks

0.93

Activations Density 0.036%

terms related to a person's history or past achievements, specifically focusing on their track record

No Comments

No Known Activations