INDEX

Explanations

Hebrew words or characters

oai_token-act-pair · gpt-3.5-turbo

Hebrew letters or characters

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

hell

-0.78

aston

-0.77

ophon

-0.76

alien

-0.73

arding

-0.72

kamp

-0.70

alore

-0.70

atos

-0.69

amia

-0.68

oons

-0.68

POSITIVE LOGITS

Ã±o

0.75

Ľ

0.74

BLIC

0.74

ãĥ¼ãĥĨ

0.72

 partName

0.70

Äĩ

0.70

å§«

0.66

×ľ

0.66

lda

0.66

odcast

0.66

Activations Density 0.034%

Hebrew words or characters

Hebrew letters or characters

No Comments

No Known Activations