INDEX

Explanations

verb forms associated with scoring or achieving points

oai_token-act-pair · gpt-3.5-turbo

instances of scoring in sports 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

tan

-0.70

agan

-0.67

é¾įå¥ĳå£«

-0.63

lly

-0.62

fell

-0.62

 coats

-0.62

 Breed

-0.60

 Layer

-0.58

 colonization

-0.58

hue

-0.58

POSITIVE LOGITS

card

0.91

keeper

0.91

 scored

0.88

rs

0.86

cards

0.85

 scoring

0.84

heet

0.83

ific

0.82

scoring

0.81

ificial

0.81

Activations Density 0.018%

verb forms associated with scoring or achieving points

instances of scoring in sports contexts

No Comments

No Known Activations