INDEX

Explanations

mentions of terms related to pre-existing items or concepts

oai_token-act-pair · gpt-3.5-turbo

references to established or pre-existing items, systems, or regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alion

-0.84

ertodd

-0.81

romeda

-0.79

ousse

-0.74

owitz

-0.73

ique

-0.72

olic

-0.71

ÃŃa

-0.71

aminer

-0.71

atto

-0.71

POSITIVE LOGITS

existing

0.89

IRE

0.82

 existing

0.81

entials

0.78

foundland

0.75

 liabilities

0.74

 compr

0.73

 exting

0.72

ãĤ´

0.70

Interstitial

0.70

Activations Density 0.015%

mentions of terms related to pre-existing items or concepts

references to established or pre-existing items, systems, or regulations

No Comments

No Known Activations