INDEX

Explanations

terms related to honors, awards, accolades, and mentions

oai_token-act-pair · gpt-3.5-turbo

words related to honors and accolades

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

arre

-0.76

hesive

-0.74

okin

-0.70

utterstock

-0.70

onica

-0.70

arette

-0.67

tera

-0.66

acco

-0.64

icago

-0.64

IFE

-0.63

POSITIVE LOGITS

 bestowed

1.01

 honors

0.93

 conferred

0.88

 honoured

0.87

 honour

0.85

 Honour

0.84

 accol

0.82

 medal

0.82

 Medal

0.81

 Lauder

0.77

Activations Density 0.159%

terms related to honors, awards, accolades, and mentions

words related to honors and accolades

No Comments

No Known Activations