INDEX

Explanations

names of people or places

oai_token-act-pair · gpt-3.5-turbo

words and phrases that denote entities, particularly names and locations

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

 Caps

-0.93

 Catalyst

-0.90

 Bers

-0.88

 Baird

-0.88

Das

-0.87

Amp

-0.86

 Bash

-0.85

Pug

-0.84

Cap

-0.82

ASP

-0.81

POSITIVE LOGITS

loe

0.96

walker

0.87

lo

0.83

Ho

0.82

iel

0.79

que

0.79

hello

0.78

 walking

0.77

 walked

0.76

HO

0.76

Activations Density 0.373%

names of people or places

words and phrases that denote entities, particularly names and locations

No Comments

No Known Activations