INDEX

Explanations

words related to complex or profound concepts, especially those that are difficult to comprehend or explain

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "unfathom" and its variations

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

 Stall

-0.66

 depos

-0.66

 productivity

-0.64

fighting

-0.64

cler

-0.62

 regist

-0.61

papers

-0.61

 postal

-0.61

 compuls

-0.61

 spouse

-0.61

POSITIVE LOGITS

athom

1.43

 "$:/

1.00

ulous

0.89

akra

0.89

hens

0.85

assador

0.85

ilk

0.84

istor

0.84

ogether

0.84

kefeller

0.84

Activations Density 0.012%

words related to complex or profound concepts, especially those that are difficult to comprehend or explain

occurrences of the word "unfathom" and its variations

No Comments

No Known Activations