INDEX

Explanations

statements of uncertainty or doubt

oai_token-act-pair · gpt-3.5-turbo

expressions of personal feelings and opinions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 phony

-0.80

 godd

-0.74

 Stupid

-0.71

 sucker

-0.70

 bogus

-0.70

 idiot

-0.68

 Scrib

-0.68

 Coulter

-0.67

 goose

-0.67

 Godd

-0.66

POSITIVE LOGITS

 secondly

1.15

Secondly

1.13

 Whilst

1.10

etheless

1.06

 additionally

0.99

 furthermore

0.99

 whilst

0.97

 util

0.96

Similarly

0.95

also

0.95

Activations Density 1.378%

statements of uncertainty or doubt

expressions of personal feelings and opinions

No Comments

No Known Activations