INDEX

Explanations

the letter "e" appearing in electronic contexts

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'e'

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

ãĤ¤ãĥĪ

-0.89

 Painter

-0.79

ablishment

-0.74

 Kitchen

-0.72

ãĥ´ãĤ¡

-0.70

 Anthem

-0.69

 Auditor

-0.67

abouts

-0.67

 Gibraltar

-0.66

 Authority

-0.66

POSITIVE LOGITS

tymology

1.04

arthed

1.03

ugen

0.93

ternally

0.93

gypt

0.91

ighth

0.89

agles

0.88

thereum

0.84

leanor

0.83

gging

0.83

Activations Density 0.024%

the letter "e" appearing in electronic contexts

occurrences of the letter 'e'

No Comments

No Known Activations