INDEX

Explanations

terms related to possession or association with specific entities

oai_token-act-pair · gpt-3.5-turbo

instances of the word "respective" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

bay

-0.80

cher

-0.80

arer

-0.79

lam

-0.78

avis

-0.77

trak

-0.76

ker

-0.75

chers

-0.74

uve

-0.73

wood

-0.70

POSITIVE LOGITS

 respective

0.95

 opinions

0.83

 strengths

0.82

 genders

0.82

 administrations

0.82

 selves

0.81

 merits

0.79

 oppos

0.78

 colours

0.77

 newsp

0.77

Activations Density 0.010%

terms related to possession or association with specific entities

instances of the word "respective" in various contexts

No Comments

No Known Activations