INDEX

Explanations

mentions of years or time references

oai_token-act-pair · gpt-3.5-turbo

references to specific timeframes and the concept of being a first or second-year in various contexts

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

achine

-0.82

oother

-0.78

è¦ļéĨĴ

-0.77

sterdam

-0.75

UTF

-0.73

hots

-0.73

amera

-0.73

hess

-0.69

uid

-0.69

kil

-0.67

POSITIVE LOGITS

tenance

0.71

 affairs

0.70

 tack

0.67

Cre

0.66

 earners

0.65

 imposition

0.65

cession

0.64

 survival

0.63

Hold

0.63

fter

0.62

Activations Density 0.084%

mentions of years or time references

references to specific timeframes and the concept of being a first or second-year in various contexts

No Comments

No Known Activations