INDEX

Explanations

processes involving assembling or connecting objects together physically

oai_token-act-pair · gpt-3.5-turbo

actions related to physical restraint or binding

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

rium

-0.74

 Remastered

-0.69

rien

-0.68

 Plaza

-0.67

 Carbuncle

-0.67

 Scotia

-0.67

 Pathfinder

-0.66

ordan

-0.65

stadt

-0.64

 Trails

-0.63

POSITIVE LOGITS

lock

0.92

down

0.91

worms

0.85

meat

0.83

stick

0.81

worm

0.79

 down

0.78

 choke

0.78

sticks

0.78

coat

0.78

Activations Density 0.237%

processes involving assembling or connecting objects together physically

actions related to physical restraint or binding

No Known Activations