INDEX

Explanations

special characters and encoding sequences rather than meaningful text

oai_token-act-pair · gpt-3.5-turbo

sequences of symbols or patterns that appear to be related to data formatting or encoding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 destro

-0.84

ascus

-0.84

 insult

-0.73

theless

-0.72

 shot

-0.71

 shroud

-0.71

 assassins

-0.70

 choked

-0.69

 therap

-0.69

 Seym

-0.68

POSITIVE LOGITS

AppData

1.17

(\

1.05

sq

0.97

SpaceEngineers

0.93

bryce

0.93

framework

0.92

gradient

0.90

circ

0.89

root

0.89

Roaming

0.89

Activations Density 0.005%

special characters and encoding sequences rather than meaningful text

sequences of symbols or patterns that appear to be related to data formatting or encoding

No Comments

No Known Activations