INDEX

Explanations

terms related to maximum values or limits

oai_token-act-pair · gpt-3.5-turbo

references to limits or maximums in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Guys

-0.77

FORE

-0.77

 Harbour

-0.73

 Debor

-0.71

ä¸ī

-0.70

killer

-0.69

fixes

-0.68

çīĪ

-0.67

ttes

-0.64

 Bride

-0.64

POSITIVE LOGITS

 allowable

1.30

 permissible

1.09

 likelihood

1.08

 achievable

1.07

imize

1.03

 practicable

1.01

 amount

0.99

 extent

0.98

 allowed

0.97

 possible

0.97

Activations Density 0.022%

terms related to maximum values or limits

references to limits or maximums in various contexts

No Comments

No Known Activations