INDEX

Explanations

symbols specific to a particular alphabet

oai_token-act-pair · gpt-3.5-turbo

occurrences of the character 'Ŀ'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 agre

-0.85

 explan

-0.85

 ende

-0.84

 distingu

-0.82

 incorpor

-0.80

chnology

-0.80

 horizont

-0.79

 maturity

-0.78

 disse

-0.78

 behavi

-0.77

POSITIVE LOGITS

ãĤ±

0.96

¤

0.94

ãĥ¼ãĥ

0.93

ļ

0.93

ãĥŃ

0.92

ef

0.92

é¾į

0.91

irect

0.90

CPU

0.87

ãĥĥãĥī

0.85

Activations Density 0.037%

symbols specific to a particular alphabet

occurrences of the character 'Ŀ'

No Comments

No Known Activations