INDEX

Explanations

similarities or connections between different entities or situations, often involving questioning or challenging circumstances

oai_token-act-pair · gpt-3.5-turbo

recurring themes or concepts within different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

tiny

-0.63

cknowled

-0.61

 Uriel

-0.58

cknow

-0.58

activated

-0.56

CAL

-0.56

 Signs

-0.56

Gau

-0.55

hen

-0.55

LAR

-0.55

POSITIVE LOGITS

same

0.72

nings

0.72

ulative

0.70

":"/

0.67

vier

0.66

rant

0.66

kefeller

0.65

ivan

0.64

roman

0.63

iatus

0.62

Activations Density 0.144%

similarities or connections between different entities or situations, often involving questioning or challenging circumstances

recurring themes or concepts within different contexts

No Comments

No Known Activations