INDEX

Explanations

quotations or reported speech with high intensity or significance, often indicating emphasis or importance

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, particularly quotes

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

ilyn

-0.79

oid

-0.76

AGES

-0.76

ible

-0.73

aque

-0.71

areth

-0.70

oids

-0.70

ages

-0.68

enzie

-0.68

stakes

-0.68

POSITIVE LOGITS

cipled

0.76

Page

0.67

mington

0.65

 indec

0.62

bably

0.61

backer

0.61

!--

0.61

Ã°

0.60

 Zoro

0.60

onga

0.59

Activations Density 0.016%

quotations or reported speech with high intensity or significance, often indicating emphasis or importance

punctuation marks, particularly quotes

No Comments

No Known Activations