INDEX

Explanations

terms related to a specific entity or brand, potentially with variations in spelling

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "En" across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

DragonMagazine

-0.89

ĳå£«

-0.88

 Heights

-0.76

sburgh

-0.72

 Parables

-0.70

 Dare

-0.68

Ĥª

-0.68

ħĭ

-0.68

 Shant

-0.63

TPPStreamerBot

-0.63

POSITIVE LOGITS

viron

1.50

abling

1.07

vironments

1.02

forcer

1.02

rollment

1.00

zyme

1.00

velop

0.98

chanted

0.98

emy

0.97

closure

0.97

Activations Density 0.007%

terms related to a specific entity or brand, potentially with variations in spelling

mentions of the word "En" across various contexts

No Comments

No Known Activations