INDEX

Explanations

phrases related to commands or instructions

oai_token-act-pair · gpt-3.5-turbo

conversational phrases that express conditions or suggestions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Demand

-0.76

ichen

-0.74

ãĤ¼ãĤ¦ãĤ¹

-0.72

Ö¼

-0.69

hab

-0.69

hoe

-0.68

Attempts

-0.68

leneck

-0.67

moil

-0.67

sbm

-0.67

POSITIVE LOGITS

 congr

1.47

 congratulations

1.31

 remember

1.16

 beware

1.15

 apologies

1.08

 sorry

1.07

 note

1.06

 check

1.04

 thank

1.04

hey

1.04

Activations Density 0.202%

phrases related to commands or instructions

conversational phrases that express conditions or suggestions

No Comments

No Known Activations