INDEX

Explanations

references to partial information or actions

oai_token-act-pair · gpt-3.5-turbo

references to partial aspects or incomplete elements of various subjects

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

Trend

-0.94

ħĭ

-0.94

Â¯Â¯

-0.82

yers

-0.80

sers

-0.80

eers

-0.78

ORGE

-0.74

åĤ

-0.72

lished

-0.70

 Trend

-0.70

POSITIVE LOGITS

 eclipse

1.00

ity

0.98

 agon

0.74

iple

0.70

 overlapping

0.70

 recess

0.69

leasing

0.67

 overlap

0.67

ibility

0.67

 explanation

0.67

Activations Density 0.012%

references to partial information or actions

references to partial aspects or incomplete elements of various subjects

No Comments

No Known Activations