INDEX

Explanations

phrases containing the word "in" followed by some contextual information

oai_token-act-pair · gpt-3.5-turbo

phrases that include the word "in" and its context within different situations

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

insula

-0.84

ornia

-0.80

ļéĨĴ

-0.78

rete

-0.75

escent

-0.74

çļ

-0.72

emale

-0.70

onomy

-0.70

irm

-0.69

opol

-0.69

POSITIVE LOGITS

 nonetheless

1.34

etheless

1.03

 nevertheless

1.01

 reality

0.91

 hindsight

0.77

 deeper

0.75

 truth

0.74

])

0.74

 alas

0.74

 actual

0.73

Activations Density 0.267%

phrases containing the word "in" followed by some contextual information

phrases that include the word "in" and its context within different situations

No Comments

No Known Activations