INDEX

Explanations

binary choices or preferences

oai_token-act-pair · gpt-3.5-turbo

yes/no responses or conditions related to compatibility and assessments in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

uyomi

-0.67

":[

-0.65

 wiping

-0.63

 easing

-0.60

 centrif

-0.59

 alike

-0.58

isconsin

-0.58

 bathroom

-0.58

 advance

-0.57

\":

-0.57

POSITIVE LOGITS

Unknown

0.80

 Retrieved

0.67

 Requires

0.66

unknown

0.63

 Converted

0.63

 Unknown

0.62

tek

0.61

ixel

0.61

ructose

0.59

NetMessage

0.59

Activations Density 0.264%

binary choices or preferences

yes/no responses or conditions related to compatibility and assessments in various contexts

No Comments

No Known Activations