INDEX

Explanations

dates and historical references

oai_token-act-pair · gpt-3.5-turbo

phrases indicating misfortune or legacy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 dstg

-0.82

 Crazy

-0.75

HUD

-0.73

à¼

-0.73

panic

-0.73

eto

-0.72

Pokemon

-0.70

 Joker

-0.70

è¦ļéĨĴ

-0.69

 blamed

-0.69

POSITIVE LOGITS

 scholarly

1.32

 methodological

1.11

 publication

1.11

 scholars

1.09

 publishing

1.06

 literary

1.03

 publish

1.03

 venerable

1.02

 breadth

0.98

Publisher

0.98

Activations Density 1.255%

dates and historical references

phrases indicating misfortune or legacy

No Comments

No Known Activations