INDEX

Explanations

words related to nationality and ethnicity

oai_token-act-pair · gpt-3.5-turbo

conjunctions and articles frequently used in list-like contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ynthesis

-0.92

perty

-0.85

omnia

-0.81

ortex

-0.80

utenberg

-0.80

uilding

-0.77

arenthood

-0.73

umbnails

-0.72

etheus

-0.71

orescence

-0.71

POSITIVE LOGITS

 Italian

1.64

 Portuguese

1.60

 Spanish

1.54

 Spani

1.51

 Italians

1.50

 German

1.50

 Vietnamese

1.48

 Filipino

1.47

 Belgian

1.46

Italian

1.45

Activations Density 0.161%

words related to nationality and ethnicity

conjunctions and articles frequently used in list-like contexts

No Comments

No Known Activations