INDEX

Explanations

phrases indicating something is not a strict requirement or direct consequence

oai_token-act-pair · gpt-3.5-turbo

phrases indicating non-absolute or conditional statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 thanks

-0.64

 Flags

-0.64

umber

-0.63

Spr

-0.62

-0.60

 Lovely

-0.58

 Pars

-0.58

Tun

-0.57

 Thanks

-0.57

 Vector

-0.56

POSITIVE LOGITS

 necessarily

3.41

necess

1.63

 inherently

1.60

 invariably

1.26

 inevitably

1.24

 ordinarily

1.16

 materially

1.10

always

1.09

 always

1.09

 unavoid

1.09

Activations Density 0.016%

phrases indicating something is not a strict requirement or direct consequence

phrases indicating non-absolute or conditional statements

No Comments

No Known Activations