INDEX

Explanations

phrases related to upward progress or advancement

oai_token-act-pair · gpt-3.5-turbo

references to specific components or elements in a situation or context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

enced

-0.80

Ã©r

-0.78

mag

-0.77

erson

-0.77

SPONSORED

-0.76

elson

-0.76

open

-0.76

cott

-0.74

avers

-0.73

âĪ

-0.73

POSITIVE LOGITS

 stairs

0.81

 seams

0.81

 defenses

0.77

 ante

0.73

 menus

0.70

 unbeliev

0.70

 horizont

0.68

 basics

0.68

 stink

0.68

 defences

0.68

Activations Density 0.292%

phrases related to upward progress or advancement

references to specific components or elements in a situation or context

No Comments

No Known Activations