INDEX

Explanations

information related to scoring goals in various sports

oai_token-act-pair · gpt-3.5-turbo

references to goals scored 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

forums

-0.81

 Layer

-0.73

é¾įå¥ĳå£«

-0.71

 pedd

-0.67

edia

-0.64

worthiness

-0.63

fell

-0.61

 breed

-0.60

tan

-0.60

MFT

-0.60

POSITIVE LOGITS

 touchdowns

0.85

 enough

0.80

 plenty

0.76

 twice

0.75

 scored

0.74

cards

0.73

 knockout

0.70

keeper

0.70

letcher

0.69

bage

0.69

Activations Density 0.041%

information related to scoring goals in various sports

references to goals scored in sports contexts

No Comments

No Known Activations