INDEX

Explanations

adjectives describing a distinct or specific quality

oai_token-act-pair · gpt-3.5-turbo

references to specific qualities or attributes associated with things or experiences

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

agents

-0.86

apers

-0.82

asks

-0.80

measures

-0.78

osponsors

-0.77

lations

-0.76

 Worse

-0.75

sites

-0.75

 Cases

-0.75

ãĥĥãĥī

-0.74

POSITIVE LOGITS

 vibe

1.48

 aura

1.41

 intimacy

1.33

 sense

1.32

 uniqueness

1.26

 elegance

1.25

 familiarity

1.25

 urgency

1.24

 feeling

1.23

 immedi

1.23

Activations Density 0.371%

adjectives describing a distinct or specific quality

references to specific qualities or attributes associated with things or experiences

No Comments

No Known Activations