INDEX

Explanations

arguments presented in various contexts

oai_token-act-pair · gpt-3.5-turbo

terms related to debates and discussions about various arguments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

chron

-0.77

redit

-0.68

 limb

-0.64

 deed

-0.63

giene

-0.61

cy

-0.61

 sight

-0.61

 livest

-0.60

 Pixie

-0.59

kin

-0.58

POSITIVE LOGITS

 arguments

1.47

 argument

1.11

argument

1.05

 Argument

0.94

uments

0.81

 excuses

0.78

 arguing

0.78

Arg

0.78

 debates

0.77

Override

0.77

Activations Density 0.007%

arguments presented in various contexts

terms related to debates and discussions about various arguments

No Comments

No Known Activations