INDEX

Explanations

mentions of the word "Na" followed by a number, possibly related to names, titles, or measurements

oai_token-act-pair · gpt-3.5-turbo

references to the chemical element sodium

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

eering

-0.80

âĶģ

-0.77

*/(

-0.74

displayText

-0.73

 guiActiveUnfocused

-0.72

tips

-0.72

dress

-0.67

ãĥī

-0.67

ttes

-0.66

hops

-0.66

POSITIVE LOGITS

Na

1.13

Na

1.02

elson

1.00

Ã¯ve

0.92

Äį

0.92

Ni

0.91

itri

0.85

emic

0.84

cipled

0.83

emon

0.82

Activations Density 0.003%

mentions of the word "Na" followed by a number, possibly related to names, titles, or measurements

references to the chemical element sodium

No Comments

No Known Activations