INDEX

Explanations

the word "Hyatt" at different levels of emphasis and detail

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific brand or entity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

mint

-0.90

*/(

-0.77

éĹĺ

-0.77

ãĤ´ãĥ³

-0.76

tumblr

-0.76

ivity

-0.72

ional

-0.70

 Showdown

-0.66

++++++++++++++++

-0.65

ificantly

-0.65

POSITIVE LOGITS

Hy

1.01

giene

0.98

Hy

0.94

brids

0.89

undai

0.88

tymology

0.82

ulia

0.81

hy

0.80

awei

0.80

rule

0.79

Activations Density 0.008%

the word "Hyatt" at different levels of emphasis and detail

mentions of a specific brand or entity

No Comments

No Known Activations