INDEX

Explanations

negative critiques or challenges

oai_token-act-pair · gpt-3.5-turbo

references to problems, criticisms, and challenges

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

++;

-0.65

liv

-0.64

ãģĨ

-0.58

zig

-0.57

NetMessage

-0.56

Know

-0.56

$.

-0.56

ãĥ´ãĤ¡

-0.55

majority

-0.53

eteria

-0.53

POSITIVE LOGITS

 arises

1.10

 relates

1.07

 revolves

1.05

is

0.99

 involves

0.96

 occurs

0.86

 derives

0.83

 seems

0.83

 comes

0.82

 surrounds

0.81

Activations Density 0.128%

negative critiques or challenges

references to problems, criticisms, and challenges

No Comments

No Known Activations