INDEX

Explanations

stock photos

oai_token-act-pair · gpt-3.5-turbo

references to stock photography and related content

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

BLE

-0.78

 Antar

-0.71

undai

-0.71

urally

-0.70

Äĩ

-0.69

 nonexistent

-0.69

 unfamiliar

-0.68

ĨĴ

-0.67

sole

-0.67

 phased

-0.66

POSITIVE LOGITS

holm

1.10

yards

1.08

ings

1.04

holders

1.01

yard

0.97

photo

0.97

holder

0.91

dale

0.90

utherland

0.88

still

0.88

Activations Density 0.015%

stock photos

references to stock photography and related content

No Comments

No Known Activations