INDEX

Explanations

phrases related to asserting or demonstrating something to be true

oai_token-act-pair · gpt-3.5-turbo

phrases related to proving something, particularly in contexts of innocence or correctness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

etting

-0.71

ussen

-0.68

artment

-0.66

ussed

-0.64

 Variant

-0.63

isions

-0.61

osp

-0.60

 livest

-0.60

hner

-0.59

erton

-0.58

POSITIVE LOGITS

 convinc

0.90

 otherwise

0.85

 innocence

0.82

 correctness

0.82

 pudding

0.80

 beyond

0.80

 validity

0.78

 definitively

0.78

 causation

0.77

 doub

0.75

Activations Density 0.081%

phrases related to asserting or demonstrating something to be true

phrases related to proving something, particularly in contexts of innocence or correctness

No Comments

No Known Activations