INDEX

Explanations

words related to online forums and threads

oai_token-act-pair · gpt-3.5-turbo

references to online forums and discussions

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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.72

lys

-0.71

stroke

-0.69

displayText

-0.69

emb

-0.63

agar

-0.63

 Siem

-0.63

ograp

-0.62

Attorney

-0.60

bat

-0.60

POSITIVE LOGITS

 thread

1.31

 moderators

1.23

 threads

1.19

 moderator

1.15

 forums

1.11

 forum

1.10

 discussion

1.06

 Forums

1.04

 Thread

1.03

 moder

1.02

Activations Density 0.090%

words related to online forums and threads

references to online forums and discussions

No Comments

No Known Activations