INDEX

Explanations

phrases indicating a purpose or intention, particularly using the word "to."

oai_token-act-pair · gpt-3.5-turbo

phrases indicating purpose or intention

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Advertisement

-0.69

 Transcript

-0.65

 drafts

-0.64

TION

-0.62

 Required

-0.61

ahime

-0.61

 disproportionate

-0.60

Already

-0.60

 Description

-0.60

 copied

-0.60

POSITIVE LOGITS

 celebrate

1.15

othy

1.04

 teach

0.95

 help

0.94

ilet

0.92

 commemorate

0.92

 defend

0.92

 promote

0.89

ulhu

0.89

bernatorial

0.88

Activations Density 0.144%

phrases indicating a purpose or intention, particularly using the word "to."

phrases indicating purpose or intention

No Comments

No Known Activations