INDEX

Explanations

words related to improvements

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "improved" 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 @ 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enter

-0.69

eps

-0.65

Unknown

-0.63

 Rite

-0.61

Witness

-0.61

 Masquerade

-0.60

root

-0.60

 spokes

-0.59

Ann

-0.59

 Cull

-0.59

POSITIVE LOGITS

 improved

3.53

 Improved

2.27

Improved

2.17

 improvement

2.07

 improving

2.06

 improvements

2.05

impro

2.03

 improves

1.93

 enhanced

1.90

 improve

1.85

Activations Density 0.017%

words related to improvements

occurrences of the term "improved" in various contexts

No Comments

No Known Activations