INDEX

Explanations

technical terms and procedures related to various fields and processes

oai_token-act-pair · gpt-3.5-turbo

mentions of techniques and methodologies 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 @ 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

deen

-0.71

 Newsp

-0.55

white

-0.53

isha

-0.53

lings

-0.53

vin

-0.52

 Syndrome

-0.52

 PRESIDENT

-0.51

 ILCS

-0.51

inas

-0.51

POSITIVE LOGITS

 adaptive

0.68

 targeting

0.68

 devised

0.68

 designed

0.64

irection

0.64

designed

0.63

 applied

0.63

 aimed

0.62

oufl

0.61

 implemented

0.61

Activations Density 1.148%

technical terms and procedures related to various fields and processes

mentions of techniques and methodologies in various contexts

No Comments

No Known Activations