INDEX

Explanations

numerical values, specifically numbers from a specific sequence

oai_token-act-pair · gpt-3.5-turbo

special characters or symbols in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

nesota

-0.91

afort

-0.82

NetMessage

-0.82

essing

-0.82

otle

-0.79

anol

-0.77

ograp

-0.76

isner

-0.76

zona

-0.75

ilic

-0.75

POSITIVE LOGITS

³

0.96

¡

0.90

ł

0.86

»

0.84

ãĥ¼ãĥ

0.83

ãĤª

0.83

´

0.82

²

0.77

ĸ

0.76

¹

0.75

Activations Density 0.008%

numerical values, specifically numbers from a specific sequence

special characters or symbols in the text

No Comments

No Known Activations