INDEX

Explanations

words related to official reports or documents

oai_token-act-pair · gpt-3.5-turbo

references to reports and their contents 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 @ 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

conservancy

-0.64

east

-0.63

 spac

-0.61

 bene

-0.59

 actionGroup

-0.58

eas

-0.57

 mortar

-0.57

ansk

-0.57

awe

-0.57

 customs

-0.56

POSITIVE LOGITS

 detailing

0.95

 titled

0.95

 commissioned

0.91

orial

0.90

velop

0.90

card

0.88

 compiled

0.85

 issued

0.83

 outlining

0.83

 documenting

0.83

Activations Density 0.071%

words related to official reports or documents

references to reports and their contents in various contexts

No Comments

No Known Activations