INDEX

Explanations

specific date and event mentions in news articles

oai_token-act-pair · gpt-3.5-turbo

repeated phrases or constructs with similar patterns or structures

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

 sacrific

-0.87

 pyramid

-0.73

 persist

-0.71

 dupl

-0.70

esp

-0.70

 ruth

-0.69

 cooper

-0.68

 weave

-0.67

 tangled

-0.67

 unchecked

-0.67

POSITIVE LOGITS

iversary

1.52

October

1.15

January

1.13

December

1.13

November

1.12

ï¸ı

1.09

February

1.09

morning

1.08

April

1.08

day

1.07

Activations Density 0.148%

specific date and event mentions in news articles

repeated phrases or constructs with similar patterns or structures

No Comments

No Known Activations