INDEX

Explanations

instances where the word "write" is used

oai_token-act-pair · gpt-3.5-turbo

instructions or references related to writing activity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Ĭ±

-0.88

agara

-0.74

ella

-0.69

EGA

-0.68

nels

-0.64

allows

-0.64

rolet

-0.62

nel

-0.61

azar

-0.61

liv

-0.61

POSITIVE LOGITS

 poems

0.93

 poetry

0.91

itatively

0.91

acters

0.87

 essays

0.85

 screenplay

0.84

 aloud

0.82

 letters

0.80

 memoir

0.79

 memos

0.77

Activations Density 0.058%

instances where the word "write" is used

instructions or references related to writing activity

No Comments

No Known Activations