INDEX

Explanations

words related to torches or torch-like objects

oai_token-act-pair · gpt-3.5-turbo

references to torches

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Austral

-0.94

ranch

-0.73

Finish

-0.73

title

-0.71

 Programme

-0.70

groups

-0.70

Americ

-0.68

safe

-0.67

omm

-0.66

 reconc

-0.66

POSITIVE LOGITS

 torches

1.90

 Gork

0.73

 sails

0.67

 Geral

0.67

ãĥķãĤ©

0.66

rolog

0.63

 Cutter

0.63

 Roose

0.63

 Raphael

0.63

 Stevens

0.63

Activations Density 0.001%

words related to torches or torch-like objects

references to torches

No Comments

No Known Activations