INDEX

Explanations

references to achievements and personal actions by individuals

oai_token-act-pair · gpt-3.5-turbo

mentions of award-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

.","

-0.78

kefeller

-0.72

 [âĢ¦]

-0.69

.""

-0.64

."

-0.60

 domestically

-0.60

."[

-0.59

itary

-0.57

*.

-0.57

 [...]

-0.56

POSITIVE LOGITS

 Packers

0.72

iannopoulos

0.72

 Pacers

0.69

CFL

0.68

 Jarrett

0.67

 scrimmage

0.64

 Bengals

0.63

 Manziel

0.63

 Redskins

0.61

 hugs

0.61

Activations Density 1.962%

references to achievements and personal actions by individuals

mentions of award-related topics

No Comments

No Known Activations