INDEX

Explanations

words related to physical modifications or transformations, especially those related to the suffix "ed"

oai_token-act-pair · gpt-3.5-turbo

past tense verbs

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

roups

-0.66

pects

-0.65

pec

-0.64

ocobo

-0.64

Battery

-0.63

zbollah

-0.60

horse

-0.60

identified

-0.59

 Swap

-0.59

href

-0.58

POSITIVE LOGITS

ed

4.12

Ed

1.53

Ed

1.47

ED

1.23

eff

1.10

eb

1.03

eds

0.97

ted

0.94

edu

0.93

op

0.92

Activations Density 0.016%

words related to physical modifications or transformations, especially those related to the suffix "ed"

past tense verbs

No Comments

No Known Activations