INDEX

Explanations

terms related to programming

oai_token-act-pair · gpt-3.5-turbo

the prefix "ex" 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 @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 scaled

-0.63

 manned

-0.62

 piled

-0.61

 pave

-0.61

 lean

-0.61

 payday

-0.60

jri

-0.60

 spac

-0.59

weeney

-0.58

Ń·

-0.58

POSITIVE LOGITS

ex

4.06

EX

2.10

Ex

1.95

Ex

1.53

EX

1.40

ex

1.35

ext

1.33

examination

1.27

ec

1.19

exp

1.16

Activations Density 0.015%

terms related to programming

the prefix "ex" in various contexts

No Comments

No Known Activations