INDEX

Explanations

translations of words or phrases from different languages into English

oai_token-act-pair · gpt-3.5-turbo

references to languages or the context in which terms or names are defined

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

icides

-0.83

idays

-0.79

rift

-0.74

 spends

-0.73

LEASE

-0.70

icable

-0.70

iership

-0.70

 tirelessly

-0.70

apses

-0.69

hett

-0.69

POSITIVE LOGITS

 English

1.23

 Arabic

1.18

 english

1.17

 Sanskrit

1.15

 Latin

1.14

 Hebrew

1.13

 Spanish

1.07

 Greek

1.05

 Wikipedia

1.01

 Hindi

1.00

Activations Density 0.189%

translations of words or phrases from different languages into English

references to languages or the context in which terms or names are defined

No Comments

No Known Activations