INDEX

Explanations

specific character sequences within a mix of Hebrew and other languages

oai_token-act-pair · gpt-3.5-turbo

words or phrases in a non-standard script or encoding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

matically

-0.85

urally

-0.83

xus

-0.77

lycer

-0.74

achy

-0.73

ulously

-0.73

arest

-0.72

osaurus

-0.72

ysis

-0.72

gart

-0.72

POSITIVE LOGITS

ãģ¦

1.62

ãģ

1.53

ãĤĭ

1.52

ãģŁ

1.51

ãĤī

1.45

ãģ¾

1.44

ãģĦ

1.44

ãĤĵ

1.41

ãĤ

1.40

ãģĭ

1.35

Activations Density 0.008%

specific character sequences within a mix of Hebrew and other languages

words or phrases in a non-standard script or encoding

No Comments

No Known Activations