INDEX

Explanations

numbered or bulleted lists and hierarchical structures

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and their associated placements in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

bered

-0.64

ment

-0.63

iffe

-0.61

gered

-0.60

oud

-0.59

azo

-0.59

itarian

-0.59

Reilly

-0.59

 itself

-0.58

POSITIVE LOGITS

 Provided

0.98

How

0.93

 Alright

0.91

 http

0.90

 https

0.90

Why

0.89

 Which

0.88

 What

0.88

 Provides

0.87

 Nope

0.86

Activations Density 0.164%

numbered or bulleted lists and hierarchical structures

punctuation marks and their associated placements in text

No Comments

No Known Activations