INDEX

Explanations

instances of punctuation marks, especially commas

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Shinra

-0.66

 exha

-0.65

 Meter

-0.64

 isolate

-0.61

 bombing

-0.61

 dome

-0.60

 Rounds

-0.58

bay

-0.58

 neighbourhood

-0.57

 flank

-0.57

POSITIVE LOGITS

ï¸ı

1.18

âĢ

1.04

_-

1.02

sorry

0.91

cause

0.89

respond

0.89

Ïī

0.89

[/

0.88

Ð²

0.87

where

0.87

Activations Density 0.146%

instances of punctuation marks, especially commas

No Comments

No Known Activations