INDEX

Explanations

phrases related to specific numerical values and symbols

oai_token-act-pair · gpt-3.5-turbo

occurrences of an unusual or non-standard character representation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

wagen

-0.73

 shroud

-0.73

 detached

-0.71

 destro

-0.70

creen

-0.66

 stricken

-0.65

 Dupl

-0.64

 Mirage

-0.64

gib

-0.63

 Bris

-0.62

POSITIVE LOGITS

certain

1.01

Ĳ

1.00

ª

0.98

¹

0.97

ĳ

0.97

ł

0.97

Ĵ

0.96

once

0.96

against

0.92

oral

0.92

Activations Density 0.190%

phrases related to specific numerical values and symbols

occurrences of an unusual or non-standard character representation

No Comments

No Known Activations