INDEX

Explanations

references to specific years or time periods

oai_token-act-pair · gpt-3.5-turbo

references to time periods and events

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

 answ

-0.81

tap

-0.73

 ingred

-0.66

 lawy

-0.65

wo

-0.64

auga

-0.63

lua

-0.61

ciation

-0.61

 gorilla

-0.61

rote

-0.60

POSITIVE LOGITS

 onwards

0.88

 season

0.84

 semester

0.82

-'

0.81

 Madness

0.80

Eid

0.79

imester

0.78

 anniversary

0.76

Era

0.76

ember

0.74

Activations Density 0.516%

references to specific years or time periods

references to time periods and events

No Comments

No Known Activations