INDEX

Explanations

phrases that refer to prior instances or comparisons to earlier states or data

oai_token-act-pair · gpt-3.5-turbo

references to prior events or periods in time

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

lee

-0.84

 ILCS

-0.76

hots

-0.74

kamp

-0.73

rage

-0.73

aliation

-0.73

lua

-0.71

ocracy

-0.69

rosso

-0.68

umes

-0.67

POSITIVE LOGITS

 generations

1.08

 incarnation

1.04

 occupant

0.99

 incarn

0.91

ebin

0.90

 batch

0.89

 generation

0.88

 iteration

0.84

 editions

0.80

vernment

0.79

Activations Density 0.023%

phrases that refer to prior instances or comparisons to earlier states or data

references to prior events or periods in time

No Comments

No Known Activations