INDEX

Explanations

phrases related to investigating or searching

oai_token-act-pair · gpt-3.5-turbo

instances of the word "dig," indicating a focus on exploration or investigation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

SAY

-0.76

 Blueprint

-0.75

 ACTIONS

-0.74

BuyableInstoreAndOnline

-0.73

Thomas

-0.70

Ren

-0.69

 Mate

-0.67

................................................................

-0.67

 Sweden

-0.65

VALUE

-0.64

POSITIVE LOGITS

dig

1.43

 digs

1.15

dug

1.14

 digging

1.07

itial

0.97

imon

0.94

oxin

0.90

Dug

0.90

itals

0.89

 trenches

0.89

Activations Density 0.008%

phrases related to investigating or searching

instances of the word "dig," indicating a focus on exploration or investigation

No Comments

No Known Activations