INDEX

Explanations

world-related terms

oai_token-act-pair · gpt-3.5-turbo

references to the concept of 'world.'

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

ippi

-0.96

veyard

-0.93

EStream

-0.86

WARE

-0.83

ibility

-0.79

edin

-0.78

shapeshifter

-0.77

irez

-0.77

ients

-0.74

Acknowled

-0.74

POSITIVE LOGITS

wide

0.92

 renown

0.85

 premiere

0.82

 wide

0.82

 domination

0.79

 heavyweight

0.77

 religions

0.75

 Airways

0.74

 renowned

0.73

icz

0.73

Activations Density 0.053%

**world-related** terms

references to the concept of 'world.'

No Comments

No Known Activations

world-related terms