INDEX

Explanations

words related to a specific language or alphabet with non-English characters

oai_token-act-pair · gpt-3.5-turbo

instances of a specific character or symbol associated with a particular language or encoding

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

ength

-0.91

accompanied

-0.88

arsity

-0.87

uality

-0.81

ajo

-0.80

 athlet

-0.79

oaded

-0.78

ateral

-0.77

adium

-0.77

oppable

-0.75

POSITIVE LOGITS

·

1.32

Ð¼

1.24

ÑĢ

1.19

Ð½

1.18

Ĺ

1.16

¶

1.11

Ðº

1.10

Ð»

1.09

°

1.09

Ð

1.08

Activations Density 0.012%

words related to a specific language or alphabet with non-English characters

instances of a specific character or symbol associated with a particular language or encoding

No Comments

No Known Activations