INDEX

Explanations

phrases related to movement or direction

oai_token-act-pair · gpt-3.5-turbo

instances of the word "back" 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 @ 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

aucus

-0.59

anto

-0.58

 Disclosure

-0.57

 delegation

-0.56

iston

-0.55

 disclaim

-0.55

weather

-0.54

 Booth

-0.54

uchi

-0.54

 Attend

-0.54

POSITIVE LOGITS

packs

1.23

fired

1.17

fires

1.12

wards

1.11

stab

1.05

dated

1.04

packing

1.00

haul

0.99

tracking

0.94

doors

0.89

Activations Density 0.054%

phrases related to movement or direction

instances of the word "back" in various contexts

No Comments

No Known Activations