INDEX

Explanations

keywords related to feedback, reports, and instructions

oai_token-act-pair · gpt-3.5-turbo

references to feedback mechanisms and reports about experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

£ı

-0.68

Abstract

-0.66

senal

-0.65

];

-0.64

];

-0.64

azard

-0.61

ãĤ´ãĥ³

-0.61

 WATCHED

-0.58

Had

-0.58

anmar

-0.56

POSITIVE LOGITS

 shouldn

1.57

 will

1.41

 should

1.40

isn

1.36

 helps

1.35

 wont

1.31

 WILL

1.28

 doesn

1.21

 deserves

1.20

 doesnt

1.20

Activations Density 0.513%

keywords related to feedback, reports, and instructions

references to feedback mechanisms and reports about experiences

No Comments

No Known Activations