INDEX

Explanations

adjectives and adverbs used to intensify a statement

oai_token-act-pair · gpt-3.5-turbo

significant emotional or physical impacts described in exaggerated terms

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zman

-0.68

sson

-0.65

ammy

-0.64

 meantime

-0.63

romeda

-0.63

 Lanka

-0.63

heast

-0.62

rients

-0.62

ugal

-0.62

Same

-0.61

POSITIVE LOGITS

 practically

1.04

 scarcely

0.94

 barely

0.93

 unrecogn

0.92

 almost

0.87

 unus

0.86

 hardly

0.84

 even

0.84

almost

0.83

 downright

0.82

Activations Density 0.222%

adjectives and adverbs used to intensify a statement

significant emotional or physical impacts described in exaggerated terms

No Comments

No Known Activations