INDEX

Explanations

terms related to positive and negative correlations between different factors or variables

oai_token-act-pair · gpt-3.5-turbo

words related to positive and negative correlations or sentiments

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

iens

-0.76

 GOODMAN

-0.75

 flaws

-0.73

ocene

-0.71

enegger

-0.70

 horizont

-0.70

 loopholes

-0.70

GER

-0.69

ador

-0.68

raltar

-0.67

POSITIVE LOGITS

 impacting

1.08

 impacted

0.95

 correlated

0.90

 disposed

0.90

 charged

0.88

 affecting

0.86

 minded

0.82

 effected

0.82

 inclined

0.80

 positioned

0.79

Activations Density 0.011%

terms related to positive and negative correlations between different factors or variables

words related to positive and negative correlations or sentiments

No Comments

No Known Activations