INDEX

Explanations

descriptions of technical specifications or performance metrics

oai_token-act-pair · gpt-3.5-turbo

technical details related to computing resources and performance

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

 Ally

-0.80

 undercover

-0.73

 Holden

-0.73

Male

-0.71

neau

-0.70

Â¨

-0.70

 Consent

-0.67

perty

-0.67

 grounding

-0.67

 Stanley

-0.67

POSITIVE LOGITS

 cached

1.55

 caching

1.49

CPU

1.48

 bandwidth

1.47

 cache

1.45

 latency

1.44

 throughput

1.44

cpu

1.44

 caches

1.42

cache

1.32

Activations Density 0.460%

descriptions of technical specifications or performance metrics

technical details related to computing resources and performance

No Comments

No Known Activations