INDEX

Explanations

multi-letter acronyms that contain 'FS', particularly emphasizing those with high activation values

oai_token-act-pair · gpt-3.5-turbo

references to financial systems and regulatory bodies

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

âĹ¼

-0.87

 Verge

-0.74

tons

-0.71

 Chomsky

-0.69

 Raider

-0.64

selves

-0.63

Das

-0.62

wcs

-0.61

owsky

-0.61

butt

-0.60

POSITIVE LOGITS

ruits

0.98

ESSION

0.95

essions

0.89

DF

0.87

ession

0.86

ometimes

0.86

folios

0.85

andom

0.85

OUND

0.83

emen

0.83

Activations Density 0.010%

multi-letter acronyms that contain 'FS', particularly emphasizing those with high activation values

references to financial systems and regulatory bodies

No Comments

No Known Activations