INDEX

Explanations

references to going deep or immersion

oai_token-act-pair · gpt-3.5-turbo

instances of the word "dive" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

numbered

-0.70

urers

-0.69

Islamic

-0.69

minus

-0.68

 wart

-0.68

Serial

-0.67

Population

-0.65

Edited

-0.65

large

-0.65

Percent

-0.64

POSITIVE LOGITS

 dive

1.32

 diving

1.23

 Dive

1.19

 dives

1.16

tails

0.85

 horizont

0.78

 Diver

0.77

olicy

0.77

 digs

0.76

itals

0.75

Activations Density 0.009%

references to going deep or immersion

instances of the word "dive" in various contexts

No Comments

No Known Activations