INDEX

Explanations

names of locations, cities, and people

oai_token-act-pair · gpt-3.5-turbo

names of specific individuals or entities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pora

-1.04

 blockers

-0.79

ribune

-0.75

lighting

-0.74

ricanes

-0.72

rib

-0.70

ters

-0.69

terness

-0.68

 Pradesh

-0.68

bos

-0.67

POSITIVE LOGITS

yrics

0.87

ovie

0.79

erous

0.78

 Parm

0.71

Maz

0.69

atures

0.69

 Dodd

0.67

ATURE

0.66

FTWARE

0.66

 Muse

0.65

Activations Density 0.024%

names of locations, cities, and people

names of specific individuals or entities

No Comments

No Known Activations