INDEX

Explanations

phrases related to sitting

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "sit on" in various contexts

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

ONSORED

-0.79

Cub

-0.76

UTF

-0.72

rament

-0.72

TPP

-0.71

States

-0.69

ress

-0.69

anan

-0.68

ELF

-0.67

RL

-0.67

POSITIVE LOGITS

 behalf

0.92

shore

0.90

 shaky

0.84

 board

0.84

erous

0.83

top

0.81

etime

0.81

 Github

0.77

 GitHub

0.76

 average

0.76

Activations Density 0.160%

phrases related to sitting

instances of the phrase "sit on" in various contexts

No Comments

No Known Activations