INDEX

Explanations

numerical values indicating a specific amount or limit

oai_token-act-pair · gpt-3.5-turbo

mentions of the number "200" and related numerical representations or contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 streak

-0.77

 pronounced

-0.67

 melanch

-0.67

 interpretation

-0.66

 Stre

-0.65

Laz

-0.64

 commentary

-0.64

 opinion

-0.62

Gle

-0.60

 interpretations

-0.59

POSITIVE LOGITS

3.35

2.24

2.23

2.14

2.11

2.10

1.96

1.88

1.87

1.84

Activations Density 0.021%

numerical values indicating a specific amount or limit

mentions of the number "200" and related numerical representations or contexts

No Comments

No Known Activations