INDEX

Explanations

situations where puzzles are mentioned and potentially solved

oai_token-act-pair · gpt-3.5-turbo

references to puzzles 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

redits

-0.74

oln

-0.72

Nau

-0.68

 temperatures

-0.66

 accredited

-0.66

rity

-0.65

ublished

-0.65

glomer

-0.64

itizen

-0.64

lav

-0.64

POSITIVE LOGITS

 puzzle

1.13

 puzzles

1.08

 Puzzle

0.91

 solving

0.90

ĸļ

0.88

 Puzzles

0.83

 Puzz

0.81

 puzz

0.80

 Pieces

0.80

 solved

0.79

Activations Density 0.011%

situations where puzzles are mentioned and potentially solved

references to puzzles in various contexts

No Comments

No Known Activations