INDEX

Explanations

words related to highly valued or sought-after items or achievements

oai_token-act-pair · gpt-3.5-turbo

terms related to awards and recognition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ghan

-0.78

annis

-0.77

harm

-0.75

sil

-0.71

thy

-0.70

osit

-0.69

sky

-0.68

gans

-0.68

bats

-0.66

uber

-0.66

POSITIVE LOGITS

 coveted

0.94

 prestigious

0.92

 awards

0.88

 honors

0.85

 accol

0.82

 adolesc

0.81

IFIED

0.80

 award

0.79

 mosqu

0.77

 occas

0.76

Activations Density 0.026%

words related to highly valued or sought-after items or achievements

terms related to awards and recognition

No Comments

No Known Activations